How many people use AI regularly. The benefits only arrive when AI is part of everyday work, not an occasional experiment.
Depth of use
How far people go beyond one-off questions. Reusable instructions, working with their own files, connections and agents are where most of the time savings come from.
Collaboration
Whether people share prompts and work together in AI tools. Shared know-how spreads good practice faster and stops everyone starting from scratch.
Skills and enablement
Training, confidence and knowing what the tools can do. Confident, trained people get better results and make fewer mistakes.
Outcomes
Time saved, how much rework is needed and whether output is checked. This is the return on the time and money spent on AI.
Governance and safety
Approved tools, company accounts, policy and checks. Good controls let people use AI more freely without putting customer data or your reputation at risk.
This grade covers how your people use AI. Strategy, data and platform maturity are looked at in discovery. It is based on what people chose to report, and the people who answered may be keener on AI than those who did not.
What to do first
Where to focus first
Starting points for a conversation, not a diagnosis. Lowest scores first, with serious data risks at the top.
Governance · 49/100 · Making it consistent
Make safe use the easy option
What we found. 42% of people use AI on a personal account, and 17% of people put customer, financial or HR information into one.
A first step. Embed governance in how you operate: move everyone onto approved company accounts, assign an owner to each tool and agent, and review risk on a set cycle.
Show details
Why it matters. Information put into personal accounts leaves your control, and you cannot see or get it back.
Questions to discuss
Which tools would you be comfortable approving today?
We roll out company accounts to everyone who needs them, train people on safe use and the policy, and support the switch so people actually move off personal accounts.
AI governance
Training
Adoption
Tool roll-out
Collaboration · 25/100 · Getting started
Share what works
What we found. Most people work with AI on their own, and little is shared between colleagues.
Your own report includes a first step, questions to discuss and how Fiveways could help for this area.
Depth of use · 32/100 · Getting started
Go beyond one-off questions
What we found. 25% of people who use AI mostly ask one-off questions or reuse simple instructions.
Your own report includes a first step, questions to discuss and how Fiveways could help for this area.
Skills · 56/100 · Making it consistent
Build skills and confidence
What we found. People did not know about 23% of the features we asked about, in tools they already use.
Your own report includes a first step, questions to discuss and how Fiveways could help for this area.
Want to talk these through?
A free 30-minute call with the Fiveways team about your results and a sensible next step.
How each department scores, out of 100. You know which departments matter most for your organisation. Sensitive information on personal accounts is flagged wherever we found it.
DepartmentOverallGovernanceWeekly useSensitive info on personal accounts
Sales6 people555283%2 people
Operations6 people554883%1 person
Finance6 people454550%None
Marketing6 people425250%1 person
Departments with fewer than 5 people answering are combined into “Other departments”, so nobody can be picked out.
All six areas by department
Department
Adoption
Depth of use
Collaboration
Skills and enablement
Outcomes
Governance and safety
Sales
83
40
31
55
68
52
Operations
83
47
30
60
63
48
Finance
50
27
20
61
67
45
Marketing
50
13
17
49
69
52
What each department uses AI for
i
We asked:Your department; and for each tool, How often? and What do you use it for?
Why it matters. AI use is rarely even. Teams that are ahead can show others what works, and teams that are behind are where help pays off fastest.
We asked:For each tool, Which account do you use?, What kinds of information do you put into it?, Where does the output go? and Does a person check it first?
Why it matters. When people use AI on personal accounts, your company’s information may no longer be under your control: you cannot see it, decide who has access, or remove it when someone leaves, and it may not be covered by your contracts with customers. Unchecked AI output sent to customers can also carry mistakes in your name. Fixing these first lets people use AI more, with confidence.
Reported by people themselves and flagged for discussion. This is not a compliance assessment.
42%use AI on a personal account for at least one tool
17%put customer, financial or HR information into AI on a personal account
0%send AI output to customers without anyone checking it
29%of tool use is on tools the company has approved
Accounts used, across all tools
Each tool a person uses counts once, so this is the share of tool use, not of people.
Company account61%
Own paid account11%
Own free account29%
Figures are the share of tool uses who answered.
Tools and how deeply they are used
Tools in use
i
We asked:Which of these AI tools do you use at work?
Why it matters. Knowing which tools people really use shows where to put licences, training and controls. Several tools doing the same job usually means wasted spend, and company information spread across more places than you know.
Microsoft 365 Copilot29%
Microsoft Copilot25%
Claude13%
Perplexity13%
ChatGPT8%
Fathom8%
Notion AI8%
Gemini4%
Otter4%
Zapier4%
Figures are the share of people who answered.
How deeply people use AI
i
We asked:How far do you go with these features?; Do you use AI agents?
Why it matters. Most of the time saved comes from the higher steps: reusable instructions, working with your own files, connecting to company data and handing over whole tasks. People who stay on one-off questions get only a small part of the value of tools you may already pay for.
Each of the 16 people who use AI is placed on the highest step they have reached.
13%of people who use AI
1 AskOne-off questions
13%of people who use AI
2 RefineSaved instructions
19%of people who use AI
3 OrganiseProjects and own files
44%of people who use AI
4 ExtendConnections and skills
13%of people who use AI
5 DelegateAgents
Features people know about
A short bar means few people know the feature exists: the cheapest gap to close, with a short demonstration.
Saved instructions or custom assistants100%
Skills87%
Projects, spaces and your own files81%
Memory and personalisation87%
Connections to company data and tools94%
Deep research or web research81%
Agents and scheduled tasks56%
Shared team workspaces, prompts or assistants31%
Figures are the share of people who answered who answered.
AI agents
i
We asked:Do you use AI agents? These are tools that carry out several steps for you, or run on their own.
Why it matters. Agents can take whole multi-step jobs off people’s hands, which is where the biggest time savings are heading. They act with less supervision, so someone needs to check their work.
Yes, regularly0%
Occasionally6%
I tried them and stopped38%
No0%
Not sure56%
Figures are the share of people who answered.
Choosing a model
i
We asked:Do you choose which model to use for a task?
Why it matters. Models differ in speed, quality and cost. Using the most powerful one for simple tasks wastes money and energy, so people who choose on purpose get better answers at a sensible cost.
Yes, I pick the model to suit the task0%
Sometimes: I mostly use the default and switch for harder tasks6%
I stick with one model I’ve chosen myself0%
I leave it on the default or automatic setting56%
I didn’t know I could choose0%
I know I can, but I don’t understand the differences0%
My company’s set-up doesn’t let me choose0%
Not sure38%
Figures are the share of people who answered.
Your people
What holds people back
i
We asked:What holds you back from using AI more?
Why it matters. These are the barriers to remove. Most, such as time to learn, training or unclear rules, cost less to fix than new tools.
No time to learn33%
No training25%
Worries about data21%
Not sure where it would help17%
I’m not sure I’m allowed13%
Worries about accuracy13%
No approved tools13%
Figures are the share of people who answered.
Where time goes
i
We asked:Which repeatable tasks take up most of your time?
Why it matters. Repeatable tasks that take a lot of time are the best first places for AI, because the gain is easy to measure.
Customer queries38%
Data entry31%
Emails31%
Reviewing documents31%
Scheduling25%
Research25%
Meeting follow-up13%
Reporting6%
Figures are the share of people who answered.
What people would hand to AI
i
We asked:What would you most like to hand to AI?
Why it matters. These are the jobs people already want help with, so starting here wins support quickly.
Analysing data44%
Summarising38%
Customer communication31%
Meeting notes25%
Automating tasks25%
Drafting and writing19%
Research19%
Figures are the share of people who answered.
Next steps
Talk it through with the Fiveways team
A 30-minute call to go through your results and what a sensible next step would be. No preparation needed.
This report looks at how your people use AI in their everyday work. It does not cover the areas below, and Fiveways can help with these too.
Building your own AI tools and appsTools for your own team, and apps or products for your customers. From quick prototypes built with AI coding assistants (sometimes called vibe coding) to properly engineered, secure products, including when to build and when to buy.AI-Powered Apps
Customer engagementUsing AI to reach and engage your customers: marketing and campaign content in your voice, search (SEO) and visibility in AI answers, a website assistant that answers questions and captures enquiries, and personalised email and follow-up, all checked before it goes out.Customer Engagement
AI strategy, data and technologyWhere AI fits in your plans, whether your data is ready, which platforms to use, and how to upgrade your technology so you can make greater use of AI in future.Foundations
How this is calculated
Six areas are each scored from 0 to 100 by fixed rules: adoption, depth of use, collaboration, skills and enablement, outcomes, and governance and safety. The overall score is their average, with equal weights. Stages: Exploring 0 to 20, Experimenting 21 to 40, Structured 41 to 60, Scaling 61 to 80, Leading 81 to 100. Governance also counts the answers you gave at setup about how AI is managed. No AI is used to score. Comparisons with other organisations will appear once enough have completed the assessment. Method version m1.
About Fiveways AI
Fiveways AI is an independent AI consultancy. We help leadership teams make sure AI is used efficiently, effectively, ethically and safely within their organisation, from strategy to delivery.